Currently, automated tests are chosen from a list of files and launched by a human operator on each of the machines allocated to the effort of testing. These machines have to be supervised so when one test is finished, another one may be launched by the operator. The results of the tests have to be collected and analyzed by the operator in order to decide if a finished test needs to be re-run. As the operator is working on or with respect to one machine, one or more of the other machines may be idle.